Synovial Fluid Biomarkers Used to Predict the Degree of Articular Cartilage Injury in the Knee

Abstract

Injury-driven changes to the intra-articular microenvironment are thought to initiate the development of post-traumatic osteoarthritis. The purpose of this study was to evaluate the accuracy with which synovial fluid and clinical data can predict the presence of high-grade cartilage lesions following intra-articular injury. Synovial fluid aspirate, collected immediately prior to surgical incision from prospectively enrolled patients, was processed using a multiplex magnetic bead immunoassay to determine the concentration of 10 pre-determined cytokines and chemokines. The degree of cartilage injury was assessed intraoperatively using ICRS grading. Logistic regression models were built to predict the presence of high-grade cartilage lesions based on clinical and biomarker data. The models were validated using repeated k-fold cross-validation, and accuracy and quality were assessed using likelihood-ratio tests. 249 patients (mean age: 41.42+/-14.02 years) were enrolled. These patients either had isolated ACL injuries (N=37), isolated meniscus injuries (N=118), combined ACL-meniscus injuries (N=76), or focal chondral lesions (N=18). Four logistic regression models were compared (Table 1). The full model, with 10 synovial fluid biomarker concentrations, age, BMI, and duration of time between injury and surgery, was found to predict high-grade ICRS cartilage damage with 85.71% accuracy. The isolated-biomarker model, consisting of all 10 biomarker concentrations, was found to have a 71.43% prediction accuracy. The isolated-clinical model was found to have a 71.43% prediction accuracy. The stepwise model, consisting of 6 biomarkers and the clinical variables, was found to predict high-grade ICRS cartilage damage with 83.67% accuracy. Through likelihood ratio testing, the stepwise model was confirmed to be superior to the other models. This study demonstrates that the combination of clinical variables and 6 synovial fluid biomarkers can be used to accurately predict high-grade cartilage lesions following intra-articular injury. This suggests that there are distinct inflammatory phenotypes among different degrees of cartilage damage with potential prognostic benefit.